CN109214856A - The method for digging and device, computer equipment and readable medium that user is intended to - Google Patents
The method for digging and device, computer equipment and readable medium that user is intended to Download PDFInfo
- Publication number
- CN109214856A CN109214856A CN201810845269.8A CN201810845269A CN109214856A CN 109214856 A CN109214856 A CN 109214856A CN 201810845269 A CN201810845269 A CN 201810845269A CN 109214856 A CN109214856 A CN 109214856A
- Authority
- CN
- China
- Prior art keywords
- user
- intent
- information
- intent information
- intention
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 54
- 230000006399 behavior Effects 0.000 claims description 62
- 238000004590 computer program Methods 0.000 claims description 4
- 230000007812 deficiency Effects 0.000 abstract description 6
- 238000012549 training Methods 0.000 description 25
- 239000000284 extract Substances 0.000 description 12
- 238000009412 basement excavation Methods 0.000 description 7
- 238000000605 extraction Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 230000005291 magnetic effect Effects 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 230000003542 behavioural effect Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 241001269238 Data Species 0.000 description 2
- HUTDUHSNJYTCAR-UHFFFAOYSA-N ancymidol Chemical compound C1=CC(OC)=CC=C1C(O)(C=1C=NC=NC=1)C1CC1 HUTDUHSNJYTCAR-UHFFFAOYSA-N 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000003466 anti-cipated effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
- G06Q30/0255—Targeted advertisements based on user history
Landscapes
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Game Theory and Decision Science (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention provides a kind of method for digging and device, computer equipment and readable medium that user is intended to.Its method includes: to excavate several intent informations of each user according to the historical behavior data of user each in multiple users;According to several intent informations of each user, the intent information of each user is extended, to carry out advertisement recommendation.Technical solution of the present invention can make up for it the deficiencies in the prior art, excavate the intent information of user effectively, comprehensively, in order to which information carries out effectively advertisement recommendation according to the user's intention, recommend efficiency so as to effectively improve advertisement.
Description
[technical field]
The present invention relates to method for digging and device, meter that computer application technology more particularly to a kind of user are intended to
Calculate machine equipment and readable medium.
[background technique]
With the propulsion of time of cell-phone, content ecology flourishes, and more and more displaying class scene flows measure present people
In daily life, advertiser has a mind to obtain more displaying class business flows.
In the prior art, under information flow scene, purpose of the user primarily for browsing, no explicitly retrieval behavior,
Advertiser can not directly acquire the intention of user.Mostly just according to some characteristic informations of user such as age, occupational identity, emerging
The features such as interest hobby carry out advertisement recommendation to user.But under this scene, since the real intention of user can not be obtained,
Cause the advertisement not necessarily user recommended really to want to click the advertisement checked in browsing, recommends to imitate so as to cause advertisement
Rate is lower.
Based on the above issues, the present invention it is urgent to provide under a kind of information flow scene user be intended to excavation scheme, so as to
In being intended to progress advertisement recommendation according to user, so that further increasing advertisement recommends efficiency.
[summary of the invention]
The present invention provides method for digging and device, computer equipment and readable medium that a kind of user is intended to, for more
The deficiencies in the prior art are mended, a kind of excavation scheme that effectively user is intended to is provided.
The present invention provides a kind of method for digging that user is intended to, which comprises
According to the historical behavior data of the user each in multiple users, several intentions letter of each user is excavated
Breath;
According to several intent informations of each user, the intent information of each user is extended, to carry out advertisement recommendation.
The present invention provides a kind of excavating gear that user is intended to, and described device includes:
Module is excavated, for the historical behavior data according to the user each in multiple users, excavates each user
Several intent informations;
Expansion module extends the intent information of each user for several intent informations according to each user, with
Carry out advertisement recommendation.
The present invention also provides a kind of computer equipment, the equipment includes:
One or more processors;
Memory, for storing one or more programs;
When one or more of programs are executed by one or more of processors, so that one or more of processing
Device realizes the method for digging that user as described above is intended to.
The present invention also provides a kind of computer-readable mediums, are stored thereon with computer program, which is held by processor
The method for digging that user as described above is intended to is realized when row.
The method for digging and device, computer equipment and readable medium that user of the invention is intended to, by according to multiple use
The historical behavior data of each user in family, excavate several intent informations of each user;According to several intent informations of each user,
The intent information of each user is extended, to carry out advertisement recommendation.Technical solution of the present invention can make up for it the deficiencies in the prior art,
Effectively, the intent information of user is excavated comprehensively, in order to which information carries out effectively advertisement recommendation according to the user's intention, thus
Advertisement can be effectively improved and recommend efficiency.
[Detailed description of the invention]
Fig. 1 is the flow chart for the method for digging embodiment one that user of the invention is intended to.
Fig. 2 is the flow chart for the method for digging embodiment two that user of the invention is intended to.
Fig. 3 is a kind of relation schematic diagram of intent information and user provided in an embodiment of the present invention.
Fig. 4 is the flow chart for the method for digging embodiment three that user of the invention is intended to.
Fig. 5 is the structure chart for the excavating gear embodiment one that user of the invention is intended to.
Fig. 6 is the structure chart for the excavating gear embodiment two that user of the invention is intended to.
Fig. 7 is the structure chart of computer equipment embodiment of the invention.
Fig. 8 is a kind of exemplary diagram of computer equipment provided by the invention.
[specific embodiment]
To make the objectives, technical solutions, and advantages of the present invention clearer, right in the following with reference to the drawings and specific embodiments
The present invention is described in detail.
Fig. 1 is the flow chart for the method for digging embodiment one that user of the invention is intended to.As shown in Figure 1, the present embodiment
The method for digging that user is intended to, can specifically include following steps:
100, according to the historical behavior data of user each in multiple users, several intent informations of each user are excavated;
101, according to several intent informations of each user, the intent information of each user is extended, to carry out advertisement recommendation.
The executing subject for the method for digging that the user of the present embodiment is intended to is user intention mining device, which is intended to dig
Pick device can be set in the server, all historical behavior data of all users serviced can be got, thus base
In all historical behavior data of all users, the intent information of each user is excavated and extended.
Specifically, when the excavation that the user of the present embodiment is intended to, the history according to user each in multiple users is first had to
Behavioral data excavates several intent informations of each user.For example, in order to obtain the real-time intent information of user, the present embodiment
In, historical behavior data of each user within the historical time period of arest neighbors, the historical time week of arest neighbors can be acquired
Phase can be arranged according to demand, such as can be 8 hours, 24 hours, one week, one month or the other times of meet demand
Cycle length, it is not limited here.In addition, the historical behavior data of the present embodiment may include search when user scans for
Historical behavior data, contextual information when can also include user's browsing pages.Search history behavior number when wherein searching for
According to the intent information that can directly react user.In contextual information when user's browsing pages, each click is checked clear
The page of looking at can imply when having the tendency that user's browsing, to imply the intent information for having user.
In the present embodiment, according to the historical behavior data of user each in multiple users, several intentions of each user are excavated
When information, semantic parsing can be carried out directly from each historical behavior data of each user, obtain the intention letter of user
Breath.For example, usually being wrapped if the intent information of user, which is excavated, to be very easy to when historical behavior data are search history behavioral data
It is contained in search term.And when intent information of the page browsed based on user to excavate user, the content for needing to analyze is more,
The intent information that user can be got, cause the excavation of user intent information to take a long time, digging efficiency it is lower.
Based on this problem, in the present embodiment, can also provide when a kind of historical behavior data are the browsing pages of user
The excavation scheme of the intent information of user.In the present embodiment, an intention can be trained to extract model in advance, to be based on the meaning
Figure extracts model realization and excavates to the intention of the user in browsing pages.For example, the step 100 can specifically include at this time
Following steps:
(a1) title and crucial content of text of the browsing pages of each historical behavior data of user are obtained;
In the present embodiment, each historical behavior data of user correspond to a browsing pages, each browsing pages
It is corresponding there are a title, foundation of the title as user intention mining of the browsing pages is obtained in the present embodiment.Another party
Can also there be the crucial content of text that one section of content of text is the browsing pages in the browsing pages of face, it is clear can effectively to summarize this
Look at the content of the page.In practical application, if include abstract in browsing pages, the crucial content of text of the browsing pages can be
Abstract in browsing pages.If do not include abstract in browsing pages, the first segment based on browsing pages can generally also summarize
The content of the browsing pages is introduced, crucial content of text of the first segment as browsing pages of browsing pages can be chosen at this time.
Or other sections can also be selected according to the characteristic of browsing pages, the crucial content of text such as final stage as browsing pages.
(b1) title of browsing pages is segmented, obtains multiple title participles;
The participle of the present embodiment can use participle technique commonly used in the prior art, and the title of browsing pages is split as
Multiple title participles, can refer to related art, details are not described herein in detail.
(c1) model being extracted using content keyword trained in advance, multiple contents keys are extracted from crucial content of text
Word;
In the present embodiment, it is also necessary to train a content keyword to extract model in advance, it can be from one section of content of text
Extract the multiple content keywords that can be identified for that this section of content.Based on the principle, mould can be extracted using the content keyword
Type extracts multiple content keywords from crucial content of text.
When the content keyword of the present embodiment extracts model training, it can be carried out using the training principle of neural network model
Training.Preparatory acquisition training is needed it is anticipated that may include hundreds of thousands training data in the training expectation of such as the present embodiment, is instructed
The quantity for practicing data is more, and trained content keyword extracts the more accurate of model extraction.Each training number of the present embodiment
It may include one section of training text gathered in advance and multiple content keywords in.
When training, each training data is input to the content keyword and is extracted in model, obtained content keyword and mention
Take multiple content keywords of model extraction.Then judge multiple content keywords of the multiple content keyword languages extracted acquisition
Whether consistent, if inconsistent, modification content keyword extracts the parameter of model, so that the multiple content keyword languages acquisition extracted
Multiple content keywords it is consistent.According to above-mentioned training method, using all training datas to content keyword extraction model
It is trained, until multiple content keywords of extraction and multiple content keywords of acquisition are consistent, training is finished, and determines content
The parameter of keyword extraction model, so that it is determined that content keyword extracts model.
(d1) multiple titles participle and multiple content keywords are combined, forms the corresponding text of corresponding historical behavior data
Eigen information;
That is, the text feature information of each historical behavior data includes that multiple title participles and multiple contents are closed
Keyword, these information all most possibly identify the content of the corresponding browsing pages of corresponding historical behavior data.
(e1) intention trained according to the corresponding text feature information of each historical behavior data of user and in advance is extracted
Model obtains several intent informations of user.
For each historical behavior data of user, by its corresponding text feature information input to meaning trained in advance
Figure extracts in model, it is intended that extracts the model prediction corresponding intent information of historical behavior data.In this way, based on same user's
A plurality of historical behavior data, several intent informations of the available user.
The training principle that the intention of the present embodiment extracts model is similar to above content keyword extraction model training, needs
Acquisition training corpus in advance, similarly may include hundreds of thousands training data in training corpus, can be in each training data
Corresponding intent information including training text characteristic information and acquisition.When then training, by the instruction in each training data
Practice text feature information input to the intention to extract in model, which extracts its intent information of model prediction, then will prediction
Intent information and the intent information of acquisition compare, whether consistent both judge, if inconsistent, adjustment is intended to extract model
Parameter so that the two is consistent.According to above-mentioned training method, instructed using all training datas to extraction model is intended to
Practice, until the intent information of prediction and the intent information of acquisition are consistent, training is finished, determine the parameter for being intended to extract model, from
And it determines intention and extracts model.
It still optionally further, will if all obtain more intent information in each historical behavior data of user
The quantity that will lead to the intent information of the user finally obtained is more.It at this time can be in each historical behavior for excavating user
After the corresponding intent information of data, each intent information text feature information corresponding with this historical behavior data is judged
Semantic similarity retains the corresponding intention letter of this historical behavior data if similarity is greater than default semantic similarity threshold value
Breath, otherwise, deletes the corresponding intent information of this historical behavior data.Intention corresponding for each historical behavior data
Information carries out after as above handling, and can effectively delete the inappropriate intent information in part of user, retains more appropriate, accurate
Intent information.
After above-mentioned processing, several intent informations of each available user.But the intention of each user
The coverage area of information or limited, in the present embodiment, is also based on several intent informations of each user, extends each use
The intent information at family, to carry out advertisement recommendation.For example, the intent information for extending each user in the present embodiment step 101 specifically may be used
To include in a manner of following at least one:
Mode A: according to several intent informations of each user, the intent information of each user is extended from good friend's dimension, is obtained each
The first intention information list of user;
The intent information that each user is extended from good friend's dimension of the present embodiment, refers specifically to: if user A and user B
For good friend, then the intent information of user B can be also used as to the intent information of user A, in this way, the just intention letter of extending user A
Breath.It similarly, can also be by the intent information of user A, as the intent information of user B, to extend the intention letter of user B
Breath.According to the extension of the user intent information of above-mentioned good friend's dimension, all intent informations after the extension of available each user,
And constitute first intention information list.
Mode B: according to several intent informations of each user, the intent information of each user is extended from intent information dimension, is obtained
To the second intention information list of each user.
The slave intent information dimension of the present embodiment extends the intent information of each user, refers specifically to: if intent information X
It is similar intent information with intent information Y, and in the intent information list of user A only includes intent information X, can will anticipates at this time
Figure information Y is also used as the intent information of user A, in this way, the just intent information of extending user A.Similarly, if the intention of user B is believed
Ceasing in list only includes intent information Y, intent information X can be also used as to the intent information of user B at this time, in this way, just extending
The intent information of user B.According to the extension of the user intent information of above-mentioned intent information dimension, the extension of available each user
All intent informations afterwards, and constitute second intention information list.
In practical application, any one in aforesaid way A and mode B can choose, or can also be two kinds with simultaneous selection
Mode carries out the extension of the intent information of user.By the extension of the intent information of the user of the present embodiment, can effectively dig
Dig all intent informations of user.In this way, can be carried out effectively wide based on the intent information of user in information flow scene
It accuses and recommends, so as to effectively improve the accuracy of advertisement recommendation, and then increase the clicking rate of advertisement, really realize advertisement
Commercial intention and commercial value.
The method for digging that the user of the present embodiment is intended to, passes through the historical behavior number according to user each in multiple users
According to excavating several intent informations of each user;According to several intent informations of each user, the intent information of each user is extended, with
Carry out advertisement recommendation.The technical solution of the present embodiment can make up for it the deficiencies in the prior art, excavate user's effectively, comprehensively
Intent information pushes away in order to which information carries out effectively advertisement recommendation according to the user's intention so as to effectively improve advertisement
Recommend efficiency.
Fig. 2 is the flow chart for the method for digging embodiment two that user of the invention is intended to.As shown in Fig. 2, the present embodiment
The method for digging that user is intended to further is extended from good friend's dimension on the basis of the technical solution of above-mentioned embodiment illustrated in fig. 1
The angle of the intent information of user, introduces technical solution of the present invention in further detail.As shown in Fig. 2, the user of the present embodiment
The method for digging of intention, can specifically include following steps:
200, according to the historical behavior data of user each in multiple users, several intent informations of each user are excavated;
In the present embodiment, the specific implementation of the step can refer to the step 100 of above-mentioned embodiment illustrated in fig. 1 in detail,
This is repeated no more.
201, according to several intent informations of each user, the corresponding user list of each intent information is obtained;
According to above-mentioned steps 200, several intent informations of each available user, such as each user's
Intent information can be expressed as<user,intent>list.As shown in figure 3, for a kind of intention letter provided in an embodiment of the present invention
The relation schematic diagram of breath and user.If user A includes 2 intent informations intent1, intent2, the intention letter of the user A
Breath list can be expressed as<userA,intent1,intent2>.User B include 3 intent information intent1,
The intent information list of intent2, intent3, the user B can be expressed as <userB, intent1, intent2, intent3
>.User F includes 1 intent information intent1, and the intent information list of the user F can be expressed as <userF, intent1
>.User G includes 1 intent information intent3, and the intent information list of the user G can be expressed as <userG, intent3
>.It is right in turn, when analyzing each intent information, the corresponding user list of available each intent information, such as the left side Fig. 3
Shown in side, the corresponding user list of intent1 can be expressed as<intent1, userA, userB, userF>intent1;
The corresponding user list of intent2 can be expressed as<intent2, and userA, userB>;The corresponding correspondence of intent3
User list can be expressed as<intent3, userB, userG>.Certainly, it above are only a citing in the present embodiment,
In practical application, in the manner described above, the corresponding use of any intent information can be obtained according to several intent informations of each user
Family list.
202, according to the corresponding user list of each intent information, good friend's similarity of any two user is obtained;
It is available to arrive user and user by way of inverted index according to the corresponding user list of each intent information
Between relationship, for example, per there is the intensity on the side between two users of an intent information to add 1 jointly, so as to basis
The corresponding user list of each intent information, finds out the intersection matrix between user.Such as shown in Fig. 3, according to the left side Fig. 3
The corresponding user list of intent information, carries out the mode of inverted index, on the right of available Fig. 3 shown in user based on intention
The intersection matrix of information.Every a line of intersection matrix or each column can indicate the relation table of a user and other users
It reaches.According to the relationship expression of two users, good friend's similarity of two users can be calculated.
203, it according to good friend's similarity of any two user, obtains and is higher than default good friend with good friend's similarity of each user
The mark of the good friend user of similarity threshold;
In multiple good friends, by the calculation of above-mentioned good friend's similarity, the good of any two user can be calculated
Friendly similarity, and judge whether good friend's similarity of two users is more than or equal to preset good friend's similarity threshold, if
It is, it may be considered that the two users good friend each other.If otherwise good friend's similarity of two users is less than good friend's similarity threshold,
Then think that the two users are not good friends.It, can will wherein one if a user is the good friend of another user by judgement
Good friend user identifier of the mark of a user as another user.
204, according to the mark of the good friend user of each user, several intent informations of corresponding good friend user are extended to pair
The intent information of the user answered obtains the first intention information list of each user;
The step 201-204 of the present embodiment is the step 101 of above-mentioned embodiment illustrated in fig. 1 " according to the several of each user
Intent information extends the intent information of each user from good friend's dimension " specific implementation.
The step of according to above-described embodiment, can be by the use after the good friend's user identifier for identifying each user
The intent information of the good friend user at family extends the intent information as the user.In this way, extension after user intent information not only
Including the several intent informations for belonging to the user in itself excavated in step 200, further including that above-mentioned steps 201-203 is determining is somebody's turn to do
These intent informations of the intent information of the good friend user of user, user constitute the first intention information list of the user together.
205, N number of intent information is obtained from multiple intent informations of each user after extension, to be based on N number of intent information
Carry out advertisement recommendation.
The quantity of the intent information of same user after extending in above-described embodiment may be more, recommends to improve advertisement
Efficiency in the present embodiment, can obtain N number of intent information from multiple intent informations of each user after extension, based on N number of
Intent information carries out advertisement recommendation.The N of the present embodiment be a positive integer, can be set to according to actual needs 5,10,20 or
Other positive integers of person, no longer citing repeats one by one herein.
For example, still optionally further, the step 205 of the present embodiment can specifically include following steps:
(a2) intent information in the first intention information list of each user is ranked up;
(b2) according to the sequence after sequence, top n intent information is taken from the first intention information list after sequence.
In the present embodiment, to effectively be filtered out from multiple intent informations in the first intention information list of user
N number of intent information for being best suitable for user's true intention needs first to the intent information in the first intention information list of the user
It is ranked up.For example, the step (a2) " being ranked up to the intent information of the first intention information list of each user ", specifically may be used
To include the following steps:
(a3) according to the historical behavior data of each user and corresponding good friend user, each user and corresponding good is acquired
The corresponding advertising display number of the intent information of friendly user and ad click number;
(b3) according to the corresponding advertising display number of intent information and advertisement point of each user and corresponding good friend user
The similarity for hitting number and each user and corresponding good friend user obtains in the first intention information list of corresponding user
Each intent information sequence index;
For each user, according to the historical behavior data of the user, it can excavate in historical behavior data, it should
The corresponding advertising display number of each intent information of user and ad click number.Similarly, the good friend of the user is used
For family, it can also be excavated in the historical behavior data of good friend user, the good friend according to the historical behavior data of good friend user
The corresponding advertising display number of each intent information of user and ad click number.
Then for each user, according to the corresponding advertisement exhibition of the intent information of the user and corresponding good friend user
Show number and ad click number and in conjunction with the user calculated in above-described embodiment and the good friend of corresponding good friend user
Similarity obtains the sequence index of each intent information in the first intention information list of user;For example, the sequence of the present embodiment
Index is specifically realized using following formula:
R (i, j)=Q (i) * sqrt (D (j)) * C (j);
Wherein R (i, j) indicates the sequence index of j-th of intent information in the first intention information list of i-th of user;
Q (i) indicates the corresponding good friend's similarity of i-th of user, in which: if j-th of intent information inherently i-th user's is original
When intent information, Q (i) is the similarity of i-th user and itself, value 1;D (j) is the historical behavior number of i-th of user
The corresponding advertising display number of j-th of intent information in, C (j) are that i-th of user anticipates to j-th in historical behavior data
The number of clicks of the corresponding advertisement of figure information;
If j-th of intent information sheet as the good friend user of i-th of user intent information when, good friend's similarity be i-th
The similarity of a user and good friend user;Q (i) is the similarity of i-th user and good friend user;D (j) is going through for good friend user
The corresponding advertising display number of j-th of intent information in history behavioral data, C (j) are that good friend user is right in historical behavior data
The number of clicks of the corresponding advertisement of j-th of intent information.
(c3) in the first intention information list of each user, according to descending suitable of sequence index of each intent information
Sequence is ranked up each intent information.
Finally, according to the descending sequence of sequence index, taken from first intention information list top n intent information into
Row advertisement is recommended, to advanced optimize the accuracy of advertisement recommendation, improves the clicking rate of advertisement, enhances advertisement recommendation
Efficiency.
The method for digging that the user of the present embodiment is intended to can make up for it existing by using the technical solution of above-described embodiment
There is the deficiency of technology, excavates the intent information of user, effectively, comprehensively from good friend's dimension in order to information according to the user's intention
It carries out effectively advertisement to recommend, recommends efficiency so as to effectively improve advertisement.
Fig. 4 is the flow chart for the method for digging embodiment three that user of the invention is intended to.As shown in figure 4, the present embodiment
The method for digging that user is intended to, on the basis of the technical solution of above-mentioned embodiment illustrated in fig. 1, further from intent information dimension
The angle of the intent information of extending user, introduces technical solution of the present invention in further detail.As shown in figure 4, the present embodiment
The method for digging that user is intended to, can specifically include following steps:
300, according to the historical behavior data of user each in multiple users, several intent informations of each user are excavated;
In the present embodiment, the specific implementation of the step can refer to the step 100 of above-mentioned embodiment illustrated in fig. 1 in detail,
This is repeated no more.
301, according to several intent informations of each user, the intention similarity of any two intent information is obtained;
302, it according to the intention similarity of any two intent information, obtains and is higher than default meaning with each intent information similarity
The similar intent information of figure similarity threshold;
303, according to several intent informations of each user and the corresponding similar intent information of each intent information, by user's
The corresponding similar intent information of intent information is extended to the intent information of user, obtains the second intention information list of each user;
The step 301-303 of the present embodiment is the step 101 of above-mentioned embodiment illustrated in fig. 1 " according to the several of each user
Intent information extends the intent information of each user from intent information dimension " specific implementation.
The difference of the step 201-204 of the step 301-303 and above-mentioned embodiment illustrated in fig. 2 of the present embodiment is: this reality
Applying and calculating the intention similarity of two intent informations in example is calculated based on the intersection matrix of good friend.For example, can will be in Fig. 3
Intent all replace with user, user replaces with intent, in the way of similar inverted index, it is available each
The relationship expression based on good friend of intent information.According to the relationship expression of two intent informations, two intentions can be calculated
Between intention similarity.Similarly, according to the similar mode of embodiment illustrated in fig. 2, retain the intention of each intent information
Similarity is higher than the similar intent information for being intended to similarity threshold.And by the similar intent information of each intent information of user
It is extended to the intent information of the user, in this way, the intent information of user not only includes that itself excavated in step 300 belongs to after extension
It further include the phase of each intent information for the user that above-mentioned steps 301-302 is determined in several intent informations of the user
Like intent information, these intent informations of user constitute the second intention information list of the user together.
304, N number of intent information is obtained from multiple intent informations of each user after extension, to be based on N number of intent information
Carry out advertisement recommendation.
The quantity of the intent information of same user after extending in above-described embodiment may be more, recommends to improve advertisement
Efficiency in the present embodiment, can obtain N number of intent information from multiple intent informations of each user after extension, based on N number of
Intent information carries out advertisement recommendation.
For example, still optionally further, the step 304 of the present embodiment can specifically include following steps:
(a4) intent information in the second intention information list of each user is ranked up;
(b4) according to the sequence after sequence, top n intent information is taken from the second intention information list after sequence.
Wherein step (a4) " being ranked up to the intent information in the second intention information list of each user ", specifically can be with
Include the following steps:
(a5) according to the historical behavior data of each user, each intention letter in the second intention information list of each user is acquired
Cease the advertising display number and ad click number of corresponding each user;
(b5) according to the advertising display number and ad click number of the corresponding each user of each intent information, each use is obtained
The sequence index of each intent information in the second intention information list at family;
For example, the sequence index of the present embodiment is specifically realized in the following way:
First according to the advertising display number and ad click number of the corresponding each user of each intent information, each intention is obtained
The corresponding average advertising display number of information and average ad click number;
Then, the sequence that each intent information in the second intention information list of each user is calculated using following formula is referred to
Number: R ' (i, j)=Q ' (j) * sqrt (aveD (j)) * (aveC (j));
Wherein R ' (i, j) indicates that the sequence of j-th of intent information in the second intention information list of i-th of user refers to
Number;Q ' (j) indicates the corresponding intention similarity of j-th of intent information, in which: if inherently i-th of j-th of intent information use
When the original intent information at family, Q ' (j) indicates the similarity of corresponding j-th of intent information and itself, value 1;If j-th
When intent information is the similar intent information of the original intent information of i-th of user, Q ' (j) indicates corresponding original intent information
To the similarity of similar intent information;AveD (j) is the corresponding average advertising display number of j-th of intent information;AveC (j) is
The corresponding average ad click number of j-th of intent information.
(c5) in the second intention information list of each user, according to descending suitable of sequence index of each intent information
Sequence is ranked up each intent information.
In the present embodiment, to effectively be filtered out from multiple intent informations in the second intention information list of user
N number of intent information for being best suitable for user's true intention needs first to the intent information in the second intention information list of the user
It is ranked up.The present embodiment is unlike the sequence of above-mentioned embodiment illustrated in fig. 2: above-mentioned embodiment illustrated in fig. 2 extension is intended to
When information increased intent information be good friend's dimension, sort index in good friend user intent information advertising display number and
Ad click number is convenient for acquisition.And increased intent information is similar intent information when extending intent information in the present embodiment,
And similar intent information may be the intent information of multiple other users simultaneously, so at this point, needing to obtain phase in sequence index
Like the average value of advertising display number of the intent information in other all users and the average value of ad click number.Remaining reality
Existing principle is identical as above-mentioned embodiment illustrated in fig. 2, can participate in the record of above-mentioned embodiment illustrated in fig. 2 in detail, no longer superfluous herein
It states.
In practical application, above-mentioned Fig. 2 can be used together with the scheme of embodiment illustrated in fig. 4, be believed respectively from first intention
N number of intent information is obtained in breath list and second intention information list, the N number of intention being based respectively in first intention information list
N number of intent information in information and second intention information list carries out advertisement recommendation.
The method for digging that the user of the present embodiment is intended to can make up for it existing by using the technical solution of above-described embodiment
There is the deficiency of technology, excavates the intent information of user, effectively, comprehensively from good friend's dimension in order to information according to the user's intention
It carries out effectively advertisement to recommend, recommends efficiency so as to effectively improve advertisement.
Fig. 5 is the structure chart for the excavating gear embodiment one that user of the invention is intended to.As shown in figure 5, the present embodiment
The excavating gear that user is intended to, can specifically include:
Module 10 is excavated for the historical behavior data according to user each in multiple users, excavates several meanings of each user
Figure information;
Expansion module 11 is used to extend the meaning of each user according to the several intent informations for excavating each user that module 10 is excavated
Figure information, to carry out advertisement recommendation.
The excavating gear that the user of the present embodiment is intended to realizes the realization for the excavation that user is intended to by using above-mentioned module
Principle and technical effect are identical as the realization of above-mentioned related method embodiment, can refer to above-mentioned related method embodiment in detail
Record, details are not described herein.
Fig. 6 is the structure chart for the excavating gear embodiment two that user of the invention is intended to.As shown in fig. 6, the present embodiment
The excavating gear that user is intended to further is introduced in further detail on the basis of the technical solution of above-mentioned embodiment illustrated in fig. 5
Technical solution of the present invention.
As shown in fig. 6, the excavating gear that the user of the present embodiment is intended to, further includes:
It obtains module 12 and is used to from multiple intent informations of each user after the extension of expansion module 11 obtain N number of intention letter
Breath, to carry out advertisement recommendation based on N number of intent information.
Still optionally further, as shown in fig. 6, the excavating gear that the user of the present embodiment is intended to, expansion module 11 are specific to wrap
It includes:
Good friend's dimension expanding element 111 is used for according to the several intent informations for excavating each user that module 10 is excavated, from good
Friendly dimension extends the intent information of each user, obtains the first intention information list of each user;And/or
It is intended to dimension expanding element 112 to be used for according to the several intent informations for excavating each user that module 10 is excavated, from meaning
Figure information dimension extends the intent information of each user, obtains the second intention information list of each user.
To be including good friend's dimension expanding element 111 and intention dimension expanding element 112 simultaneously in embodiment illustrated in fig. 6
, it can also only include one of them in practical application.
Still optionally further, in the excavating gear that the user of the present embodiment is intended to, good friend's dimension expanding element 111 is specifically used
In:
According to the several intent informations for excavating each user that module 10 is excavated, the corresponding user's column of each intent information are obtained
Table;
According to the corresponding user list of each intent information, good friend's similarity of any two user is obtained;
According to good friend's similarity of any two user, obtain similar higher than default good friend to good friend's similarity of each user
Spend the mark of the good friend user of threshold value;
According to the mark of the good friend user of each user, several intent informations of corresponding good friend user are extended to corresponding
The intent information of user obtains the first intention information list of each user.
Still optionally further, it in the excavating gear that the user of the present embodiment is intended to, obtains module 12 and is specifically used for:
The intent information in the first intention information list of each user that the processing of good friend's dimension expanding element 111 is obtained into
Row sequence;
According to the sequence after sequence, top n intent information is taken from the first intention information list after sequence.
Still optionally further, it in the excavating gear that the user of the present embodiment is intended to, obtains module 12 and is specifically used for:
According to the historical behavior data of each user and corresponding good friend user, acquires each user and corresponding good friend uses
The corresponding advertising display number of the intent information at family and ad click number;
According to the corresponding advertising display number of the intent information of each user and corresponding good friend user and ad click
Good friend's similarity of several and each user and corresponding good friend user, obtain in the first intention information list of corresponding user
Each intent information sequence index;
In the first intention information list of each user, according to the descending sequence of sequence index of each intent information,
Each intent information is ranked up.
Still optionally further, it in the excavating gear that the user of the present embodiment is intended to, obtains module 12 and is specifically used for using such as
Lower formula is realized:
R (i, j)=Q (i) * sqrt (D (j)) * C (j);
Wherein R (i, j) indicates the sequence index of j-th of intent information in the first intention information list of i-th of user;
Q (i) indicates the corresponding good friend's similarity of i-th of user, in which: if j-th of intent information inherently i-th user's is original
When intent information, Q (i) is the similarity of i-th user and itself, value 1;D (j) is the historical behavior number of i-th of user
The corresponding advertising display number of j-th of intent information in, C (j) are that i-th of user anticipates to j-th in historical behavior data
The number of clicks of the corresponding advertisement of figure information;
If j-th of intent information sheet as the good friend user of i-th of user intent information when, good friend's similarity be i-th
The similarity of a user and good friend user;Q (i) is the similarity of i-th user and good friend user;D (j) is going through for good friend user
The corresponding advertising display number of j-th of intent information in history behavioral data, C (j) are that good friend user is right in historical behavior data
The number of clicks of the corresponding advertisement of j-th of intent information.
Still optionally further, in the excavating gear that the user of the present embodiment is intended to, it is intended that dimension expanding element 112 is specifically used
In:
According to the several intent informations for excavating each user that module 10 is excavated, the intention phase of any two intent information is obtained
Like degree;
According to the intention similarity of any two intent information, obtains and be higher than default intention phase with each intent information similarity
Like the similar intent information of degree threshold value;
According to several intent informations of each user and the corresponding similar intent information of each intent information, by the intention of user
The corresponding similar intent information of information is extended to the intent information of user, obtains the second intention information list of each user.
Still optionally further, it in the excavating gear that the user of the present embodiment is intended to, obtains module 12 and is specifically used for:
To be intended to dimension expanding element 112 handle intent information in the obtained second intention information list of each user into
Row sequence;
According to the sequence after sequence, top n intent information is taken from the second intention information list after sequence.
Still optionally further, it in the excavating gear that the user of the present embodiment is intended to, obtains module 12 and is specifically used for:
According to the historical behavior data of each user, each intent information pair in the second intention information list of each user is acquired
The advertising display number and ad click number of each user answered;
According to the advertising display number and ad click number of the corresponding each user of each intent information, obtain each user's
The sequence index of each intent information in second intention information list;
In the second intention information list of each user, according to the descending sequence of sequence index of each intent information,
Each intent information is ranked up.
Still optionally further, it in the excavating gear that the user of the present embodiment is intended to, obtains module 12 and is specifically used for using such as
Under type is realized:
According to the advertising display number and ad click number of the corresponding each user of each intent information, each intention letter is obtained
Cease corresponding average advertising display number and average ad click number;
The sequence index of each intent information in the second intention information list of each user: R ' is calculated using following formula
(i, j)=Q ' (j) * sqrt (aveD (j)) * (aveC (j));
Wherein R ' (i, j) indicates that the sequence of j-th of intent information in the second intention information list of i-th of user refers to
Number;Q ' (j) indicates the corresponding intention similarity of j-th of intent information, in which: if inherently i-th of j-th of intent information use
When the original intent information at family, Q ' (j) indicates the similarity of corresponding j-th of intent information and itself, value 1;If j-th
When intent information is the similar intent information of the original intent information of i-th of user, Q ' (j) indicates corresponding original intent information
To the similarity of similar intent information;AveD (j) is the corresponding average advertising display number of j-th of intent information;AveC (j) is
The corresponding average ad click number of j-th of intent information.
The excavating gear that the user of the present embodiment is intended to realizes the realization for the excavation that user is intended to by using above-mentioned module
Principle and technical effect are identical as the realization of above-mentioned related method embodiment, can refer to above-mentioned related method embodiment in detail
Record, details are not described herein.
Fig. 7 is the structure chart of computer equipment embodiment of the invention.As shown in fig. 7, the computer equipment of the present embodiment,
It include: one or more processors 30 and memory 40, memory 40 works as memory for storing one or more programs
The one or more programs stored in 40 are executed by one or more processors 30, so that one or more processors 30 are realized such as
The method for digging that figure 1 above-embodiment illustrated in fig. 4 user is intended to.To include that multiple processors 30 are in embodiment illustrated in fig. 7
Example.
For example, Fig. 8 is a kind of exemplary diagram of computer equipment provided by the invention.Fig. 8, which is shown, to be suitable for being used to realizing this
The block diagram of the exemplary computer device 12a of invention embodiment.The computer equipment 12a that Fig. 8 is shown is only an example,
Should not function to the embodiment of the present invention and use scope bring any restrictions.
As shown in figure 8, computer equipment 12a is showed in the form of universal computing device.The component of computer equipment 12a can
To include but is not limited to: one or more processor 16a, system storage 28a connect different system components (including system
Memory 28a and processor 16a) bus 18a.
Bus 18a indicates one of a few class bus structures or a variety of, including memory bus or Memory Controller,
Peripheral bus, graphics acceleration port, processor or the local bus using any bus structures in a variety of bus structures.It lifts
For example, these architectures include but is not limited to industry standard architecture (ISA) bus, microchannel architecture (MAC)
Bus, enhanced isa bus, Video Electronics Standards Association (VESA) local bus and peripheral component interconnection (PCI) bus.
Computer equipment 12a typically comprises a variety of computer system readable media.These media can be it is any can
The usable medium accessed by computer equipment 12a, including volatile and non-volatile media, moveable and immovable Jie
Matter.
System storage 28a may include the computer system readable media of form of volatile memory, such as deposit at random
Access to memory (RAM) 30a and/or cache memory 32a.Computer equipment 12a may further include it is other it is removable/
Immovable, volatile/non-volatile computer system storage medium.Only as an example, storage system 34a can be used for reading
Write immovable, non-volatile magnetic media (Fig. 8 do not show, commonly referred to as " hard disk drive ").Although being not shown in Fig. 8,
The disc driver for reading and writing to removable non-volatile magnetic disk (such as " floppy disk ") can be provided, and non-easy to moving
The CD drive that the property lost CD (such as CD-ROM, DVD-ROM or other optical mediums) is read and write.In these cases, each
Driver can be connected by one or more data media interfaces with bus 18a.System storage 28a may include at least
One program product, the program product have one group of (for example, at least one) program module, these program modules are configured to hold
The function of the above-mentioned each embodiment of Fig. 1-Fig. 6 of the row present invention.
Program with one group of (at least one) program module 42a/utility 40a, can store and deposit in such as system
In reservoir 28a, such program module 42a include --- but being not limited to --- operating system, one or more application program,
It may include the reality of network environment in other program modules and program data, each of these examples or certain combination
It is existing.Program module 42a usually executes the function and/or method in above-mentioned each embodiment of Fig. 1-Fig. 6 described in the invention.
Computer equipment 12a can also be with one or more external equipment 14a (such as keyboard, sensing equipment, display
24a etc.) communication, the equipment interacted with computer equipment 12a communication can be also enabled a user to one or more, and/or
(such as network interface card is adjusted with any equipment for enabling computer equipment 12a to be communicated with one or more of the other calculating equipment
Modulator-demodulator etc.) communication.This communication can be carried out by input/output (I/O) interface 22a.Also, computer equipment
12a can also by network adapter 20a and one or more network (such as local area network (LAN), wide area network (WAN) and/or
Public network, such as internet) communication.As shown, network adapter 20a passes through its of bus 18a and computer equipment 12a
The communication of its module.It should be understood that although not shown in the drawings, other hardware and/or software can be used in conjunction with computer equipment 12a
Module, including but not limited to: microcode, device driver, redundant processor, external disk drive array, RAID system, tape
Driver and data backup storage system etc..
Processor 16a by the program that is stored in system storage 28a of operation, thereby executing various function application and
Data processing, such as realize the method for digging that user shown in above-described embodiment is intended to.
The present invention also provides a kind of computer-readable mediums, are stored thereon with computer program, which is held by processor
The method for digging that the intelligence user as shown in above-described embodiment is intended to is realized when row.
The computer-readable medium of the present embodiment may include in the system storage 28a in above-mentioned embodiment illustrated in fig. 8
RAM30a, and/or cache memory 32a, and/or storage system 34a.
With the development of science and technology, the route of transmission of computer program is no longer limited by tangible medium, it can also be directly from net
Network downloading, or obtained using other modes.Therefore, the computer-readable medium in the present embodiment not only may include tangible
Medium can also include invisible medium.
The computer-readable medium of the present embodiment can be using any combination of one or more computer-readable media.
Computer-readable medium can be computer-readable signal media or computer readable storage medium.Computer-readable storage medium
Matter for example may be-but not limited to-system, device or the device of electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor, or
Any above combination of person.The more specific example (non exhaustive list) of computer readable storage medium includes: with one
Or the electrical connections of multiple conducting wires, portable computer diskette, hard disk, random access memory (RAM), read-only memory (ROM),
Erasable programmable read only memory (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light
Memory device, magnetic memory device or above-mentioned any appropriate combination.In this document, computer readable storage medium can
With to be any include or the tangible medium of storage program, the program can be commanded execution system, device or device use or
Person is in connection.
Computer-readable signal media may include in a base band or as carrier wave a part propagate data-signal,
Wherein carry computer-readable program code.The data-signal of this propagation can take various forms, including --- but
It is not limited to --- electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be
Any computer-readable medium other than computer readable storage medium, which can send, propagate or
Transmission is for by the use of instruction execution system, device or device or program in connection.
The program code for including on computer-readable medium can transmit with any suitable medium, including --- but it is unlimited
In --- wireless, electric wire, optical cable, RF etc. or above-mentioned any appropriate combination.
The computer for executing operation of the present invention can be write with one or more programming languages or combinations thereof
Program code, described program design language include object oriented program language-such as Java, Smalltalk, C++,
Further include conventional procedural programming language-such as " C " language or similar programming language.Program code can be with
It fully executes, partly execute on the user computer on the user computer, being executed as an independent software package, portion
Divide and partially executes or executed on a remote computer or server completely on the remote computer on the user computer.?
Be related in the situation of remote computer, remote computer can pass through the network of any kind --- including local area network (LAN) or
Wide area network (WAN)-be connected to subscriber computer, or, it may be connected to outer computer (such as mentioned using Internet service
It is connected for quotient by internet).
In several embodiments provided by the present invention, it should be understood that disclosed system, device and method can be with
It realizes by another way.For example, the apparatus embodiments described above are merely exemplary, for example, the unit
It divides, only a kind of logical function partition, there may be another division manner in actual implementation.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of hardware adds SFU software functional unit.
The above-mentioned integrated unit being realized in the form of SFU software functional unit can store and computer-readable deposit at one
In storage media.Above-mentioned SFU software functional unit is stored in a storage medium, including some instructions are used so that a computer
It is each that equipment (can be personal computer, server or the network equipment etc.) or processor (processor) execute the present invention
The part steps of embodiment the method.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (Read-
Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic or disk etc. it is various
It can store the medium of program code.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Within mind and principle, any modification, equivalent substitution, improvement and etc. done be should be included within the scope of the present invention.
Claims (24)
1. the method for digging that a kind of user is intended to, which is characterized in that the described method includes:
According to the historical behavior data of the user each in multiple users, several intent informations of each user are excavated;
According to several intent informations of each user, the intent information of each user is extended, to carry out advertisement recommendation.
2. the method according to claim 1, wherein extension is each according to several intent informations of each user
After the intent information of the user, the method also includes:
N number of intent information is obtained from multiple intent informations of each user after extension, to be based on N number of intent information
Carry out advertisement recommendation.
3. according to the method described in claim 2, it is characterized in that, several intent informations according to each user, expand
The intent information for opening up each user, specifically includes:
According to several intent informations of each user, the intent information of each user is extended from good friend's dimension, obtains each institute
State the first intention information list of user;And/or
According to several intent informations of each user, the intent information of each user is extended from intent information dimension, is obtained
The second intention information list of each user.
4. according to the method described in claim 3, it is characterized in that, according to several intent informations of each user, from good friend
Dimension extends the intent information of each user, obtains the first intention information list of each user, specifically includes:
According to several intent informations of each user, the corresponding user list of each intent information is obtained;
According to the corresponding user list of each intent information, good friend's similarity of user described in any two is obtained;
Good friend's similarity of the user according to any two obtains and is higher than default good friend with good friend's similarity of each user
The mark of the good friend user of similarity threshold;
According to the mark of the good friend user of each user, several intent informations of the corresponding good friend user are extended to pair
The intent information of the user answered obtains the first intention information list of each user.
5. according to the method described in claim 4, it is characterized in that, from multiple intent informations of each user after extension
N number of intent information is obtained, is specifically included:
Intent information in the first intention information list of each user is ranked up;
According to the sequence after sequence, top n intent information is taken from the first intention information list after sequence.
6. according to the method described in claim 5, it is characterized in that, to the first intention information list of each user
Intent information is ranked up, and is specifically included:
According to the historical behavior data of each user and the corresponding good friend user, each user and correspondence are acquired
The good friend user the corresponding advertising display number of intent information and ad click number;
According to the corresponding advertising display number of the intent information of each user and the corresponding good friend user and advertisement point
Good friend's similarity of number and each user and the corresponding good friend user is hit, the institute of the corresponding user is obtained
State the sequence index of each intent information in first intention information list;
In the first intention information list of each user, the sequence index according to each intent information is descending
Sequence, each intent information is ranked up.
7. according to the method described in claim 6, it is characterized in that, according to each user and the corresponding good friend user
The corresponding advertising display number of intent information and ad click number and each user and the corresponding good friend user
Similarity, obtain the sequence index of each intent information in the first intention information list of the corresponding user, tool
Body is realized using following formula:
R (i, j)=Q (i) * sqrt (D (j)) * C (j);
Wherein the R (i, j) indicates j-th of intention letter in the first intention information list of i-th of user
The sequence index of breath;The Q (i) indicates the corresponding good friend's similarity of i-th of user, in which: if j-th of intention is believed
When ceasing the original intent information of inherently i-th user, the Q (i) is that i-th of user is similar to itself
Degree, value 1;The D (j) is the corresponding advertisement of j-th of intent information in the historical behavior data of i-th of user
Show number, the C (j) be i-th of user in historical behavior data to the corresponding advertisement of j-th of intent information
Number of clicks;
If j-th of intent information sheet as the good friend user of i-th of user intent information when, the good friend is similar
Degree is the similarity of i-th user and the good friend user;The Q (i) is i-th of user and the good friend user
Similarity;The D (j) is the corresponding advertisement exhibition of j-th of intent information in the historical behavior data of the good friend user
Show number, the C (j) be the good friend user in historical behavior data to the corresponding advertisement of j-th of intent information
Number of clicks.
8. according to the method described in claim 3, it is characterized in that, according to several intent informations of each user, from intention
Information dimension extends the intent information of each user, obtains the second intention information list of each user, specifically includes:
According to several intent informations of each user, the intention similarity of intent information described in any two is obtained;
The intention similarity of the intent information according to any two obtains and is higher than default meaning with each intent information similarity
The similar intent information of figure similarity threshold;
According to several intent informations of each user and the corresponding similar intent information of each intent information, by the use
The corresponding similar intent information of the intent information at family is extended to the intent information of the user, obtains the of each user
Two are intended to information list.
9. according to the method described in claim 8, it is characterized in that, from multiple intent informations of each user after extension
N number of intent information is obtained, is specifically included:
Intent information in the second intention information list of each user is ranked up;
According to the sequence after sequence, top n intent information is taken from the second intention information list after sequence.
10. according to the method described in claim 9, it is characterized in that, the second intention information list to each user
In intent information be ranked up, specifically include:
According to the historical behavior data of each user, each institute in the second intention information list of each user is acquired
State the advertising display number and ad click number of the corresponding each user of intent information;
According to the advertising display number and ad click number of the corresponding each user of each intent information, each institute is obtained
State the sequence index of each intent information in the second intention information list of user;
In the second intention information list of each user, the sequence index according to each intent information is descending
Sequence, each intent information is ranked up.
11. according to the method described in claim 10, it is characterized in that, according to the corresponding each use of each intent information
The advertising display number and ad click number at family obtain each described in the second intention information list of each user
The sequence index of intent information, is specifically realized in the following way:
According to the advertising display number and ad click number of the corresponding each user of each intent information, each institute is obtained
State the corresponding average advertising display number of intent information and average ad click number;
The sequence of each intent information in the second intention information list of each user is calculated using following formula
Index: R ' (i, j)=Q ' (j) * sqrt (aveD (j)) * (aveC (j));
Wherein the R ' (i, j) indicates j-th of intention letter in the second intention information list of i-th of user
The sequence index of breath;Q ' (j) indicates the corresponding intention similarity of j-th of intent information, in which: if described in j-th
When the original intent information of inherently i-th user of intent information, Q ' (j) indicates corresponding j-th of meaning
The similarity of figure information and itself, value 1;If j-th of intent information is the original intent information of i-th of user
Similar intent information when, Q ' (j) indicates that the corresponding original intent information is similar to the similar intent information
Degree;The aveD (j) is the corresponding average advertising display number of j-th of intent information;AveC (j) is j-th of institute
State the corresponding average ad click number of intent information.
12. the excavating gear that a kind of user is intended to, which is characterized in that described device includes:
Module is excavated, for the historical behavior data according to the user each in multiple users, excavates the number of each user
A intent information;
Expansion module extends the intent information of each user for several intent informations according to each user, to carry out
Advertisement is recommended.
13. device according to claim 12, which is characterized in that described device further include:
Module is obtained, for obtaining N number of intent information from multiple intent informations of each user after extension, to be based on institute
It states N number of intent information and carries out advertisement recommendation.
14. device according to claim 13, which is characterized in that the expansion module specifically includes:
Good friend's dimension expanding element extends each use from good friend's dimension for several intent informations according to each user
The intent information at family obtains the first intention information list of each user;And/or
It is intended to dimension expanding element, for several intent informations according to each user, extends each institute from intent information dimension
The intent information for stating user obtains the second intention information list of each user.
15. device according to claim 14, which is characterized in that good friend's dimension expanding element is specifically used for:
According to several intent informations of each user, the corresponding user list of each intent information is obtained;
According to the corresponding user list of each intent information, good friend's similarity of user described in any two is obtained;
Good friend's similarity of the user according to any two obtains and is higher than default good friend with good friend's similarity of each user
The mark of the good friend user of similarity threshold;
According to the mark of the good friend user of each user, several intent informations of the corresponding good friend user are extended to pair
The intent information of the user answered obtains the first intention information list of each user.
16. device according to claim 15, which is characterized in that the acquisition module is specifically used for:
Intent information in the first intention information list of each user is ranked up;
According to the sequence after sequence, top n intent information is taken from the first intention information list after sequence.
17. device according to claim 16, which is characterized in that the acquisition module is specifically used for:
According to the historical behavior data of each user and the corresponding good friend user, each user and correspondence are acquired
The good friend user the corresponding advertising display number of intent information and ad click number;
According to the corresponding advertising display number of the intent information of each user and the corresponding good friend user and advertisement point
Good friend's similarity of number and each user and the corresponding good friend user is hit, the institute of the corresponding user is obtained
State the sequence index of each intent information in first intention information list;
In the first intention information list of each user, the sequence index according to each intent information is descending
Sequence, each intent information is ranked up.
18. device according to claim 17, which is characterized in that the acquisition module is specifically used for using following formula
To realize:
R (i, j)=Q (i) * sqrt (D (j)) * C (j);
Wherein the R (i, j) indicates j-th of intention letter in the first intention information list of i-th of user
The sequence index of breath;The Q (i) indicates the corresponding good friend's similarity of i-th of user, in which: if j-th of intention is believed
When ceasing the original intent information of inherently i-th user, the Q (i) is that i-th of user is similar to itself
Degree, value 1;The D (j) is the corresponding advertisement of j-th of intent information in the historical behavior data of i-th of user
Show number, the C (j) be i-th of user in historical behavior data to the corresponding advertisement of j-th of intent information
Number of clicks;
If j-th of intent information sheet as the good friend user of i-th of user intent information when, the good friend is similar
Degree is the similarity of i-th user and the good friend user;The Q (i) is i-th of user and the good friend user
Similarity;The D (j) is the corresponding advertisement exhibition of j-th of intent information in the historical behavior data of the good friend user
Show number, the C (j) be the good friend user in historical behavior data to the corresponding advertisement of j-th of intent information
Number of clicks.
19. device according to claim 14, which is characterized in that the intention dimension expanding element is specifically used for:
According to several intent informations of each user, the intention similarity of intent information described in any two is obtained;
The intention similarity of the intent information according to any two obtains and is higher than default meaning with each intent information similarity
The similar intent information of figure similarity threshold;
According to several intent informations of each user and the corresponding similar intent information of each intent information, by the use
The corresponding similar intent information of the intent information at family is extended to the intent information of the user, obtains the of each user
Two are intended to information list.
20. device according to claim 19, which is characterized in that the acquisition module is specifically used for:
Intent information in the second intention information list of each user is ranked up;
According to the sequence after sequence, top n intent information is taken from the second intention information list after sequence.
21. device according to claim 20, which is characterized in that the acquisition module is specifically used for:
According to the historical behavior data of each user, each institute in the second intention information list of each user is acquired
State the advertising display number and ad click number of the corresponding each user of intent information;
According to the advertising display number and ad click number of the corresponding each user of each intent information, each institute is obtained
State the sequence index of each intent information in the second intention information list of user;
In the second intention information list of each user, the sequence index according to each intent information is descending
Sequence, each intent information is ranked up.
22. device according to claim 21, which is characterized in that the acquisition module is specifically used in the following way
To realize:
According to the advertising display number and ad click number of the corresponding each user of each intent information, each institute is obtained
State the corresponding average advertising display number of intent information and average ad click number;
The sequence of each intent information in the second intention information list of each user is calculated using following formula
Index: R ' (i, j)=Q ' (j) * sqrt (aveD (j)) * (aveC (j));
Wherein the R ' (i, j) indicates j-th of intention letter in the second intention information list of i-th of user
The sequence index of breath;Q ' (j) indicates the corresponding intention similarity of j-th of intent information, in which: if described in j-th
When the original intent information of inherently i-th user of intent information, Q ' (j) indicates corresponding j-th of meaning
The similarity of figure information and itself, value 1;If j-th of intent information is the original intent information of i-th of user
Similar intent information when, Q ' (j) indicates that the corresponding original intent information is similar to the similar intent information
Degree;The aveD (j) is the corresponding average advertising display number of j-th of intent information;AveC (j) is j-th of institute
State the corresponding average ad click number of intent information.
23. a kind of computer equipment, which is characterized in that the equipment includes:
One or more processors;
Memory, for storing one or more programs;
When one or more of programs are executed by one or more of processors, so that one or more of processors are real
The now method as described in any in claim 1-11.
24. a kind of computer-readable medium, is stored thereon with computer program, which is characterized in that the program is executed by processor
Method of the Shi Shixian as described in any in claim 1-11.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810845269.8A CN109214856A (en) | 2018-07-27 | 2018-07-27 | The method for digging and device, computer equipment and readable medium that user is intended to |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810845269.8A CN109214856A (en) | 2018-07-27 | 2018-07-27 | The method for digging and device, computer equipment and readable medium that user is intended to |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109214856A true CN109214856A (en) | 2019-01-15 |
Family
ID=64990300
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810845269.8A Pending CN109214856A (en) | 2018-07-27 | 2018-07-27 | The method for digging and device, computer equipment and readable medium that user is intended to |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109214856A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110851729A (en) * | 2019-11-19 | 2020-02-28 | 深圳前海微众银行股份有限公司 | Resource information recommendation method, device, equipment and computer storage medium |
CN113034167A (en) * | 2019-12-24 | 2021-06-25 | 上海佳投互联网技术集团有限公司 | User interest analysis method and advertisement delivery method based on user behaviors |
CN114020795A (en) * | 2021-10-14 | 2022-02-08 | 深圳华云信息***有限公司 | Business processing method and device, electronic equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104679743A (en) * | 2013-11-26 | 2015-06-03 | 阿里巴巴集团控股有限公司 | Method and device for determining preference model of user |
CN105677780A (en) * | 2014-12-31 | 2016-06-15 | Tcl集团股份有限公司 | Scalable user intent mining method and system thereof |
CN106682925A (en) * | 2015-11-06 | 2017-05-17 | 北京奇虎科技有限公司 | Method and device for recommending advertisement content |
CN107203530A (en) * | 2016-03-16 | 2017-09-26 | 北大方正集团有限公司 | Information recommendation method |
-
2018
- 2018-07-27 CN CN201810845269.8A patent/CN109214856A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104679743A (en) * | 2013-11-26 | 2015-06-03 | 阿里巴巴集团控股有限公司 | Method and device for determining preference model of user |
CN105677780A (en) * | 2014-12-31 | 2016-06-15 | Tcl集团股份有限公司 | Scalable user intent mining method and system thereof |
CN106682925A (en) * | 2015-11-06 | 2017-05-17 | 北京奇虎科技有限公司 | Method and device for recommending advertisement content |
CN107203530A (en) * | 2016-03-16 | 2017-09-26 | 北大方正集团有限公司 | Information recommendation method |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110851729A (en) * | 2019-11-19 | 2020-02-28 | 深圳前海微众银行股份有限公司 | Resource information recommendation method, device, equipment and computer storage medium |
CN113034167A (en) * | 2019-12-24 | 2021-06-25 | 上海佳投互联网技术集团有限公司 | User interest analysis method and advertisement delivery method based on user behaviors |
CN114020795A (en) * | 2021-10-14 | 2022-02-08 | 深圳华云信息***有限公司 | Business processing method and device, electronic equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109087135A (en) | The method for digging and device, computer equipment and readable medium that user is intended to | |
CN105117474B (en) | The method and apparatus of recommendation information load are carried out in the reading model of webpage | |
CN105426514B (en) | Personalized mobile application APP recommended method | |
US10606897B2 (en) | Aggregating personalized suggestions from multiple sources | |
CN109214417A (en) | The method for digging and device, computer equipment and readable medium that user is intended to | |
CN107122400B (en) | Method, computing system and storage medium for refining query results using visual cues | |
CN109214418A (en) | The method for digging and device, computer equipment and readable medium that user is intended to | |
CN104836720A (en) | Method for performing information recommendation in interactive communication, and device | |
CN110083688A (en) | Search result recalls method, apparatus, server and storage medium | |
CN109993583B (en) | Information pushing method and device, storage medium and electronic device | |
US9171045B2 (en) | Recommending queries according to mapping of query communities | |
CN105095231A (en) | Method and device for presenting search result | |
CN109214856A (en) | The method for digging and device, computer equipment and readable medium that user is intended to | |
Ma et al. | Sketch-based annotation and visualization in video authoring | |
CN108509405A (en) | A kind of generation method of PowerPoint, device and equipment | |
US11019012B2 (en) | File sending in instant messaging application | |
CN108228556A (en) | Key phrase extracting method and device | |
Wang et al. | Similarity-based visualization of large image collections | |
CN110799971A (en) | Generation and presentation of media content | |
CN113254711A (en) | Interactive image display method and device, computer equipment and storage medium | |
CN109376219A (en) | Matching process, device, electronic equipment and the storage medium of text attributes field | |
CN110110218B (en) | Identity association method and terminal | |
CN109241323A (en) | The method for generating user's poster is commented on based on e-book and calculates equipment | |
CN105069034A (en) | Recommendation information generation method and apparatus | |
CN108108450A (en) | The method and relevant device of image procossing |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190115 |
|
RJ01 | Rejection of invention patent application after publication |